Изображения страниц
PDF
EPUB

at 110 volts with reference to its various periods of life is shown in the following table:

[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small]

The initial increase of illuminating value and of current consumed is doubtless caused by a change in the structure of the tantalum wire, this change being accompanied by a reduction. of resistance and, consequently, of the phenomena resulting therefrom. We may say at once that after a certain amount of use the filament presents a radical change in appearance when viewed with the naked eye. While the fresh filament has a perfectly smooth and cylindrical surface, it acquires a peculiarly glistening aspect as it grows old, so that a lamp having served for some time can be readily distinguished from a new lamp. When

FIG. 5.-Tantalum filament,
before and after 1,000
hours' use.

FIG. 6.-Filament frame of a new lamp.

looked at under the microscope, the filament that has burned for a length of time shows a clear tendency toward contraction and formation of drops or beads. Figure 5 is an illustration of a piece of filament in its fresh state and of the same piece after 1,000 hours of service, the specimen in each case being magnified one hundred times. This gradual shortening of the filament can also be observed in the lamps themselves, and offers a further indication of the age of a lamp. Figure 6 represents the filament frame of a new lamp. It will be noticed that the tantalum wire is led up and down and hangs loose on the supporting frame in easy wide arches, without sharp bends.

SM 1905-13

But after being used for some time the aspect of the lamp is quite different. As shown in figure 7, the wire has contracted, the wide arches have disappeared and sharp-pointed angles have taken their place.

The behavior of these lamps is most peculiar when the filament has burnt through. While with all other incandescent lamps the burning through of the filament is tantamount to the economical death of the lamp, it may happen with tantalum lamps that they burn through several times without being rendered useless; on the contrary, each burning through is followed by an increase, often considerable, of the illuminating power. This peculiar result is due to the fact that in many cases a broken wire comes in contact with its neighbor, so that the circuit is again established. A part

[graphic][subsumed][merged small][merged small]

of the filament is thus cut out of the circuit, and the lamp consequently burns more intensely, and sometimes even too intensely, in which case, of course, only a short span of life is left to it. Yet we have had more than one lamp under observation, the filament of which broke for a first time after a short period of service and then broke repeatedly, but notwithstanding this the lamp lived more than 1,000 hours. We have often succeeded in rendering a lamp with a broken filament serviceable again by tapping it to bring the broken piece into contact with its neighbor. Figure 8 represents the frame of a lamp in which the filament was burnt through in three places, and yet continued to do service. For the sake of clearness, the back spans of the filament have been omitted in the drawing, while the front spans which were carrying the current are drawn in specially heavy lines.

It must further be mentioned that after serving for some time, say 200 to 300 hours, the tantalum filament loses a great deal of its mechanical resistance, while, as has been stated by Doctor Von Bolton, tantalum wire, when new, has a greater tensile strength than steel. It becomes brittle and will break easily in the course of its life as a filament. It is therefore advisable when lamps have served for some time not to remove them from their old fittings and put them into new ones, as that might easily cause the filament to break. New lamps are not very sensitive to strong shocks, even while burning, but when this alteration in the filament has occurred it is well to preserve them from shocks.

200 Resistance

180

160

140

120

Carbon

The behavior of the tantalum lamp under a very great increase of voltage is of special interest to the incandescent-lamp maker. As was to be expected, the trials. made in this respect have also shown the superiority of this lamp over the carbon lamp. It has been ascertained that tantalum lamps for 110 volts, 25 Hefner candlepower and 15 watts per candlepower only burn through at 260 to 300 volts if the pressure is increased slowly and gradually, while with carbon lanips designed to work under the same conditions nothing like that figure can be obtained. The superiority of the tantalum lamp over the carbon lamp with regard to blackening of the glass

100

80

60

Tantalum

40

20

20

40 60 80 100 120 140 160 180 200 Voltage

FIG. 9. Variation of resistance with voltage of tantalum as compared with carbon.

globe can also be proved in a few hours by means of comparative burning tests at about 30 per cent overload.

Another advantage of the tantalum lamp over the carbon lamp is that the resistance of tantalum, like that of all other metals, strongly increases with the rise of temperature, while carbon is known to diminish in resistance when it is hot. In figure 9 the variation of the resistance of tantalum and of carbon as a function of the voltage is graphically represented, the pressure being assumed as 100 volts and the resistance at 100 arbitrary units when the efficiency is 1·5 watts per Hefner candlepower, so that for each per cent of variation of voltage the respective percentage of variation of resistance is shown. It will be seen in the first instance that the resistance of the tantalum increases to more than five times its original value from the cold state to 15 watts per Hefner candlepower, while the resistance of the carbon decreases to about one-half of its initial value. It will

further be noticed that even afterwards the resistance of tantalum goes on rising, while the resistance of carbon keeps dropping. Therefore the increase or decrease of pressure causes the strength of current, and with it the illuminating value, to rise or fall at a quicker rate in the carbon lamp than in the tantalum lamp, and, consequently, the latter is less sensitive to variations of pressure than the former.

Having thus related the whole history of the development of the tantalum lamp and fully entered into a critical comparison between it and the carbon filament lamp, we need scarcely add that we do not intend, of course, to be satisfied with what we have already obtained. For the time being, however, and until a larger building has been erected for the production of tantalum, our firm has resolved to keep to the type for which there is an immediate practical demand. That is the lamp for 100 to 120 volts, which supplies 25 Hefner candlepower at 110 volts, or will have a higher or lower illuminating value if worked at correspondingly higher or lower voltages. In conclusion, I would recapitulate the properties which we claim as peculiar characteristics of our invention as follows:

1. The tantalum lamp has a filament made of a metallic conductor, and burns at once on being connected without any previous heating.

2. The light-giving wire is prepared by melting in a vacuum and drawing. It is tough even in the cold state, and can therefore be coiled and fixed in the lamp when cold.

3. A relatively great length of wire can be placed in a simple manner within a bulb of ordinary dimensions.

4. Tantalum ore exists in considerable quantities and can be easily procured.

5. Similar principles of treatment can be adhibited to other metals of a very high melting point.

SOME REFINEMENTS OF MECHANICAL SCIENCE."

By AMRBOSE SWASEY, Cleveland, Ohio.

As we open this, the twenty-fifth annual meeting of the American Society of Mechanical Engineers, the history of the society for a quarter of a century comes before us, and it is an occasion when it is especially appropriate to make some mention of the growth and progress of the society since it was organized.

At the beginning of the society who would have dared to predict the wonderful advance that has been made in mechanical engineering. There was indeed a great field for work for just such a society. The long list of meetings which have been so fully attended and so valuable to the members; the transactions, with their records of addresses, papers, and discussions by men of experience in nearly every branch of mechanical engineering, and the constant growth of the society until at the present time it has a membership of nearly 2,900, all go to show that from the beginning it has been an earnest and progressive organization, and a most important factor in the progress of mechanical science and of the mechanic arts.

Not only those of us who were counted among its first members, but those who from year to year have been added to its membership, may well feel proud of its splendid record.

The scope and influence of the society, which has been constantly increasing in the past, will surely continue, and never was its future brighter than at present.

For the subject of my address I wish to speak of a few of those methods and mechanisms which have been developed and perfected to such a degree of refinement that they may be considered as almost. beyond the practical, and yet were it not for such refinements they could not possibly be made to serve the utilitarian purposes which make them of such inestimable value to us all.

The division and the measurement of time is to-day, as it has been for ages, among the most important of the subjects affecting the welfare of mankind, and as time has rolled on and there has been a better

a President's address, American Society of Mechanical Engineers, December 6, 1904. Reprinted from author's corrected copy.

« ПредыдущаяПродолжить »